support multiple configurations per regression test

[vogler]

We should have a way to check the same regression test with multiple configurations to avoid duplicating files with just different PARAM: ... lines at the top.
Would be good to run several solvers on the same test. E.g. https://github.com/goblint/analyzer/blob/master/tests/regression/34-localization/01-nested.c should now also be tested with td3 after cf11c4b.
Also see #196 (comment).

We could just have several lines starting with PARAM: or something like PARAM+: which appends/overrides the previous line.

[sim642]

We could just have several lines starting with PARAM:

Indeed, I've thought about this many times. Theoretically it should be an easy change: just repeat the same test file with each PARAM given. Where is gets complicated is how to report the results of such multi-param tests: if one of the configurations passes but another fails, then it should be clear which ones fail (not that just some failed).
Also it complicates the regtest.sh script, which then must have a way to choose the configuration you want to run.

What's worse is that it would be desirable to specify expected results differently for the different configurations. Otherwise one must set them based on the most imprecise configuration and use // TODO in places where some configuration could be more precise.

[vogler]

Good points. I guess it mostly makes sense for tests where the expected results are the same (e.g. groups of solvers).
Or we come up with some extra syntax to reference the configuration -- probably better to name them, just the index might be too cryptic.

For reporting we could diff the param lines to identify which config failed (and sth like: 1 out of 2 configurations failed: ...).